Skip to content

Latest commit

 

History

History
243 lines (212 loc) · 13.1 KB

README.md

File metadata and controls

243 lines (212 loc) · 13.1 KB
Raw

Event Store

Event Store is an open-source, functional database with Complex Event Processing in JavaScript.

TL;DR;

helm repo add eventstore https://eventstore.github.io/EventStore.Charts
helm repo update

The default username and password for the admin interface is admin:changeit.

Introduction

This chart bootstraps a Event Store deployment on a Kubernetes cluster using the Helm package manager.

Prerequisites

  • Kubernetes 1.4+ with Beta APIs enabled
  • PV provisioner support in the underlying infrastructure (Only when persisting data)

Installing the Chart

Add the Event Store Charts repository.

helm repo add eventstore https://eventstore.github.io/EventStore.Charts
helm repo update

To install the Event Store chart with the release name eventstore:

helm install -n eventstore eventstore/eventstore

To install the Event Store chart with a custom admin password:

helm install -n eventstore eventstore/eventstore --set 'admin.password=<your admin password>'

This triggers Helm to run a post-install Job which resets the admin password using the Event Store HTTP API. You can then use the username admin and the password set in the above command to log into the admin interface.

The above commands install Event Store with the default configuration. The configuration section below lists the parameters that can be configured during installation.

Deleting the Chart

Delete the eventstore release.

helm delete eventstore --purge

This command removes all the Kubernetes components associated with the chart and deletes the release.

Configuration

The following table lists the configurable parameters of the Event Store chart and their default values.

Parameter Description Default
image Container image name eventstore/eventstore
imageTag Container image tag release-4.1.1-hotfix1
imagePullPolicy Container pull policy IfNotPresent
imagePullSecrets Specify image pull secrets nil
clusterSize The number of nodes in the cluster 3
admin.jobImage Post install Job image with curl installed for setting admin password tutum/curl
admin.jobImageTag Post install Job image tag latest
admin.password Custom password for admin interface (should be set in separate file) nil
admin.serviceType Service type for the admin interface ClusterIP
admin.proxyImage NGINX image for admin interface proxy nginx
admin.proxyImageTag NGINX image tag latest
podDisruptionBudget.enabled Enable a pod disruption budget for nodes false
podDisruptionBudget.minAvailable Number of pods that must still be available after eviction 2
podDisruptionBudget.maxUnavailable Number of pods that can be unavailable after eviction nil
extIp External IP address 0.0.0.0
intHttpPort Internal HTTP port 2112
extHttpPort External HTTP port 2113
intTcpPort Internal TCP port 1112
extTcpPort External TCP port 1113
gossipAllowedDiffMs The amount of drift, in ms, between clocks on nodes before gossip is rejected 600000
eventStoreConfig Additional Event Store parameters {}
scavenging.enabled Enable the scavenging CronJob for all nodes false
scavenging.image The image to use for the scavenging CronJob lachlanevenson/k8s-kubectl
scavenging.imageTag The image tag use for the scavenging CronJob latest
scavenging.schedule The schedule to use for the scavenging CronJob 0 2 * * *
persistence.enabled Enable persistence using PVC false
persistence.existingClaim Provide an existing PVC nil
persistence.accessMode Access Mode for PVC ReadWriteOnce
persistence.size Size of data volume 8Gi
persistence.mountPath Mount path of data volume /var/lib/eventstore
persistence.annotations Annotations for PVC {}
resources CPU/Memory resource request/limits Memory: 256Mi, CPU: 100m
nodeSelector Node labels for pod assignment {}
podAnnotations Pod annotations {}
tolerations Toleration labels for pod assignment []
affinity Affinity settings for pod assignment {}

Specify each parameter using the --set key=value[,key=value] argument to helm install or create a values.yaml file and use helm install --values values.yaml.

Scaling Persistent Volume

After running Event Store for a while you may run into the situation where you have outgrown the initial volume. Below will walk you through the steps to check the disk usage and how to update if necessary.

You can check the disk usage of the StatefulSet pods using the df command:

kubectl exec eventstore-0 df
Filesystem     1K-blocks     Used Available Use% Mounted on
overlay         28056816 11530904  15326860  43% /
tmpfs              65536        0     65536   0% /dev
tmpfs            7976680        0   7976680   0% /sys/fs/cgroup
/dev/nvme0n1p9  28056816 11530904  15326860  43% /etc/hosts
shm                65536        4     65532   1% /dev/shm
/dev/nvme1n1     8191416  2602160   5572872  32% /var/lib/eventstore                        --> PVC usage
tmpfs            7976680       12   7976668   1% /run/secrets/kubernetes.io/serviceaccount
tmpfs            7976680        0   7976680   0% /sys/firmware

If the Use% for /var/lib/eventstore mount is at an unacceptably high number, then follow one of the two options below depending on how your cluster is set up.

Option 1: Resize PVC created with volume expansion enabled

You can check if volume expansion is enabled on the PVC StorageClass by running:

kubectl get storageclass <pvc storageclass> -o yaml

You should see the following in the output if volume expansion is enabled:

apiVersion: storage.k8s.io/v1
kind: StorageClass
...
allowVolumeExpansion: true
...
  1. First resize the PVC for each StatefulSet pod. 
    kubectl edit pvc data-eventstore-0

    This will open up the specification in your default text editor.

    ...
spec:
accessModes:
- ReadWriteOnce
resources:
    requests:
    storage: 8Gi  --> change this to desired size
...

    Save and close the file after editing. If you get the error only dynamically provisioned pvc can be resized and the storageclass that provisions the pvc must support resize then the storage class has not enabled volume expansion. No worries! Skip down to Option 2.

  2. Delete the StatefulSet but keep the pods. 
    kubectl delete sts --cascade=false eventstore
  3. Update the chart values with the new storage request value that you edited in step (1). 
    helm upgrade eventstore eventstore/eventstore --set 'persistence.size=<value from step (1)>'

Option 2: Resize PVC created without volume expansion enabled

This process is a bit involved but also a good exercise for backing up the database. We will use AWS S3 as the storage backend but the process works just as well for other backends such as GCS.

  1. Connect to one of the StatefulSet pods. 
    kubectl exec -it eventstore-0 sh
  2. Install Python (required for AWS CLI). 
    apt-get update
apt-get install python3
export PATH="$PATH:/usr/local/bin"

    ref: https://docs.aws.amazon.com/cli/latest/userguide/install-linux-python.html

  3. Install pip. 
    curl -O https://bootstrap.pypa.io/get-pip.py
python3 get-pip.py
  4. Install the AWS CLI. 
    pip install awscli

    ref: https://docs.aws.amazon.com/cli/latest/userguide/install-linux.html

  5. Configure the AWS CLI. 
    aws configure

    Enter your credentials when prompted.

  6. Dry run the copy. 
    aws s3 cp --recursive /var/lib/eventstore/ s3://<bucket>/backup/eventstore/20190830/ --dryrun

    Change the S3 path to your preferred destination. If the copy operation looks good, proceed to the next step.

  7. Copy the files. 
    aws s3 cp --recursive /var/lib/eventstore/ s3://<bucket>/backup/eventstore/20190830/

    ref: https://eventstore.org/docs/server/database-backup/#backing-up-a-database

  8. Create a new Event Store cluster with the new volume size. It is recommended to set allowVolumeExpansion: true on your cluster's StorageClass prior to creating the cluster. This will make it easier to resize in the future by following the steps in Option 1 above. See the documentation for more details.
  9. Repeat steps (1) through (5) on the new cluster StatefulSet pod.
  10. Copy the backup files from S3. 
    aws s3 cp s3://<bucket>/backup/eventstore/20190830/chaser.chk /var/lib/eventstore/truncate.chk
aws s3 cp --recursive --exclude="truncate.chk" s3://<bucket>/backup/eventstore/20190830 /var/lib/eventstore

    ref: https://eventstore.org/docs/server/database-backup/#restoring-a-database

  11. Restart the StatefulSet pods. 
    kubectl delete $(kubectl get pod -o name -l app.kubernetes.io/component=database,app.kubernetes.io/instance=eventstore)
  12. Check the logs to ensure Event Store is processing the chunks. 
    kubectl logs -f eventstore-0
...
[00001,12,11:16:47.264] CACHED TFChunk #1-1 (chunk-000001.000000) in 00:00:00.0000495.
[00001,12,11:17:07.681] Completing data chunk 1-1...
...

Additional Resources